Evaluating the sustainability of cropping systems (CS) involves accounting for several criteria relative to social, economic and environmental domains. The identification of these criteria and their hierarchy might be highly subjective since the various stakeholders (farmers, development workers, researchers, policy makers...) often have different interests or priorities. Two types of multi-criteria analysis methods are proposed for the evaluation of CS. In the Optimization Under Multiple Constraints (OUMC), the farmer takes his decisions related to the CS according to the whole functioning of the farm i.e. the interrelations with the livestock system and the satisfaction of the family needs. Objectives or criteria relative to social or environmental aspects can be incorporated as constraints defined by other stakeholders or fixed by the producer or as goals in the utility function (multiple goal programming). On the other hand, the Multi Attribute Hierarchy (MAH) approach offers the possibility of directly incorporating a wide set of criteria covering all the aspects of sustainability. Selection, importance and order of aggregation of the criteria may be defined a priori or through a participatory process. Both techniques offer the possibility of testing the impact of changes in the economic environment of the farms, adjustments of the new cropping systems in terms of labour and cash requirement across seasons or agronomic performances, and thus not only to evaluate cropping systems but also to assist scientists in designing new ones. In a case study in the Cerrado region of Central Brazil, we used OUMC and a participatory MAH for evaluating the sustainability of new cropping systems. This paper presents the first lessons learned on the comparative advantages of these two approaches.